Introducing Pluggable Protocols


The Microsoft .NET Framework provides a layered, extensible, and managed implementation of Internet services that can be integrated quickly and easily into your applications. The Internet access classes in the System.Net and System.Net.Sockets namespaces can be used to implement both Web-based and Internet-based applications.

Internet applications can be classified broadly into two kinds: client applications that request information and server applications that respond to information requests from clients. The classic Internet client-server application is the World Wide Web, where people use browsers to access documents and other data stored on Web servers worldwide.

Applications are not limited to just one of these roles; for instance, the familiar middle-tier application server responds to requests from clients by requesting data from another server, in which case it is acting as both a server and a client.

The client application makes a request by identifying the requested Internet resource and the communication protocol to use for the request and response. If necessary, the client also provides any additional data required to complete the request, such as proxy location or authentication information (user name, password, and so on). Once the request is formed, the request can be sent to the server.

The .NET Framework uses a Uniform Resource Identifier (URI) to identify the requested Internet resource and communication protocol. The URI consists of at least three, and possibly four, fragments: the scheme identifier, which identifies the communications protocol for the request and response; the server identifier, which consists of either a Domain Name System (DNS) host name or a TCP address that uniquely identifies the server on the Internet; the path identifier, which locates the requested information on the server; and an optional query string, which passes information from the client to the server. For example, the URI "" consists of the scheme identifier "http", the server identifier "", the path "/whatsnew.aspx", and the query string "?date=today".

After the server has received the request and processed the response, it returns the response to the client application. The response includes supplemental information, such as the type of the content (raw text or XML data, for example).

The .NET Framework uses specific classes to provide the three pieces of information required to access Internet resources through a request/response model: the Uri class, which contains the URI of the Internet resource you are seeking; the WebRequest class, which contains a request for the resource; and the WebResponse class, which provides a container for the incoming response.

Client applications create WebRequest instances by passing the URI of the network resource to the Create method. This static method creates a WebRequest for a specific protocol, such as HTTP. The WebRequest that is returned provides access to properties that control both the request to the server and access to the data stream that is sent when the request is made. The GetResponse method on the WebRequest sends the request from the client application to the server identified in the URI. In cases in which the response might be delayed, the request can be made asynchronously using the BeginGetResponse method on the WebRequest, and the response can be returned at a later time using the EndGetResponse method.

The GetResponse and EndGetResponse methods return a WebResponse that provides access to the data returned by the server. Because this data is provided to the requesting application as a stream by the GetResponseStream method, it can be used in an application anywhere data streams are used.

The WebRequest and WebResponse classes are the basis of pluggable protocols — an implementation of network services that enables you to develop applications that use Internet resources without worrying about the specific details of the protocol that each resource uses. Descendant classes of WebRequest are registered with the WebRequest class to manage the details of making the actual connections to Internet resources.

As an example, the HttpWebRequest class manages the details of connecting to an Internet resource using HTTP. By default, when the WebRequest.Create method encounters a URI that begins with "http:" or "https:" (the protocol identifiers for HTTP and secure HTTP), the WebRequest that is returned can be used as is, or it can be typecast to HttpWebRequest to access protocol-specific properties. In most cases, the WebRequest provides all the necessary information for making a request.

Any protocol that can be represented as a request/response transaction can be used in a WebRequest. You can derive protocol-specific classes from WebRequest and WebResponse and then register them for use by the application with the static WebRequest.RegisterPrefix method.

When client authorization for Internet requests is required, the Credentials property of the WebRequest supplies the necessary credentials. These credentials can be a simple name/password pair for basic HTTP or digest authentication, or a name/password/domain set for NTLM or Kerberos authentication. One set of credentials can be stored in a NetworkCredentials instance, or multiple sets can be stored simultaneously in a CredentialCache instance. The CredentialCache uses the URI of the request and the authentication scheme that the server supports to determine which credentials to send to the server.

For applications that need to make simple requests for Internet resources, the WebClient class provides common methods for uploading data to or downloading data from an Internet server. WebClient relies on the WebRequest class to provide access to Internet resources; therefore, the WebClient class can use any registered pluggable protocol.

For applications that cannot use the request/response model, or for applications that need to listen on the network as well as send requests, the System.Net.Sockets namespace provides the TCPClient, TCPListener, and UDPClient classes. These classes handle the details of making connections using different transport protocols and expose the network connection to the application as a stream.

Developers familiar with the Windows Sockets interface or those who need the control provided by programming at the socket level will find that the System.Net.Sockets classes meet their needs. The System.Net.Sockets classes are a transition point from managed to native code within the System.Net classes. In most cases, System.Net.Sockets classes marshal data into their Windows 32-bit counterparts, as well as handling any necessary security checks.

Programming Pluggable Protocols
Network Programming in the .NET Framework
Network Programming Samples
Networking Samples for .NET on MSDN Code Gallery